Issue 039 — The Cyber Toolchain

Bugcrowd

1 release · 2026-08-08 Changelog

Bugcrowd is a crowdsourced security testing platform connecting organizations with ethical hackers to identify and report vulnerabilities.

Bugcrowd added the ability to generate a DNS Records Report within the platform.

└──▷ WHAT SHIPPED · 1 FEATURE 1 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

DNS Records Report generation new 30

Bugcrowd now supports generating a DNS Records Report within the platform.

— Names the report but no detail on fields or access path. changelog-20260808-e39b22ba

Was this useful?

Nmap

changes since 2026-07-09 Code

Nmap - the Network Mapper. Github mirror of official SVN repository.

Nmap 7.991 adds SSL service name matching for portrules and fixes Ncat HTTP proxy hang, libssh2 CVEs, and DNS case matching.

Adds 's' service names (e.g. amt-soap-https) to NSE portrules so scripts that run on standard unencrypted services also run on their SSL-protected equivalents.
Ncat now propagates termination signals (SIGINT, SIGTERM, SIGHUP) to child processes when using -k keep-open mode, preventing orphaned children on shutdown.

2 details in Everything

Was this useful?

Strix

1 release · 2026-08-07 Notes

Open-source AI penetration testing tool to find and fix your app’s vulnerabilities.

Strix v1.5.1 broadens its vulnerability scanner's coverage of Firebase Storage rules misconfigurations.

└──▷ WHAT SHIPPED · 1 FEATURE 0 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Expanded Firebase Storage rules coverage improved 30

The vulnerability scanner now expands coverage of Firebase Storage rules for misconfiguration detection.

— No detail on which rules or mechanism added. v1.5.1

Was this useful?

GitGuardian Detection Engine

1 release · 2026-08-08 Changelog

GitGuardian Detection Engine scans code repositories to identify and alert on exposed secrets, credentials, and sensitive data in real-time.

GitGuardian's Detection Engine added four new secret detectors (OpenCode, Lovable, Supabase), tightened false-positive filtering across five existing detectors for proxy-config and presigned-URL patterns, and retired or merged three detectors including the Amadeus and GitHub OAuth token checks.

└──▷ WHAT SHIPPED · 4 FEATURES 2 OF 4 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

False-positive filtering across generic and AWS detectors improved 75

Extends generic_password to ignore passwords in public proxy-subscription configs (shadowsocks/vmess/vless/trojan/hysteria URIs, Telegram MTProto links, clash/mihomo node lists); extends generic_high_entropy_secret and base64_generic_high_entropy_secret to ignore secrets whose decoded context is a public proxy-subscription config or a URL-signing token (S3 presigned parameters, HLS stream tokens, booking/redirect tokens); extends aws_iam to ignore presigned-URL parameters and GitHub avatar IDs; and extends generic_encryption_key to ignore Talend encrypted passwords (enc:system|routine.encryption.key.vN:... format).

— Names exact detectors and false-positive patterns suppressed. changelog-20260808-28734af6

02 01

Detector removals and merge: Amadeus, GitHub OAuth, PostgreSQL variant breaking 70

The Amadeus OAuth credentials detector and checker have been disabled because the Amadeus self-service API has been discontinued; the github_oauth_token detector and checker have been removed since the underlying service is no longer available; and the attached-port PostgreSQL variant detector has been merged into the Generic Database Assignment detector.

— Explains reason and scope for each removal or merge. changelog-20260808-28734af6

thinner coverage below

03 02

New detectors for OpenCode, Lovable and Supabase new 58

Adds OpenCode API Key detector for OpenCode API keys, Lovable Access Token detector for Lovable project access tokens, Lovable Git Token detector for Lovable git tokens, and a new Supabase Service Role JWT checker for Supabase service_role JWTs.

— Names each detector but gives no usage detail beyond that. changelog-20260808-28734af6

04 00

Azure Event Hub checker hardened improved 25

Hardens the Azure Event Hub checker with the same Atom-feed validation requirement.

— Thin one-line note; validation mechanism referenced but unexplained. changelog-20260808-28734af6

└──▷ BREAKING ON UPGRADE

!
The Amadeus OAuth credentials detector and checker have been disabled; the Amadeus self-service API has been discontinued.
!
The github_oauth_token detector and checker have been removed; the service for this detector is no longer available.
!
The attached-port PostgreSQL variant detector has been merged into the Generic Database Assignment detector.

Was this useful?

infisical

1 release · 2026-08-07 Notes

Infisical is the open-source platform for secrets, certificates, and privileged access management.

Infisical v0.162.18 extends gateway routing to Chef app connections and adds an inline role-change action for project members.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Gateway support for Chef app connection new 45

The Chef app connection now supports routing through Infisical gateways, enabling Chef integrations to be reached via a gateway rather than direct network access.

— Describes capability but no config or flag names given v0.162.18

02 00

Modify Role action on project member rows new 45

A new Modify Role action was added directly on project member role rows in the UI, letting admins change a member's role inline without navigating to a separate page.

— Names UI location but no exact menu path or API v0.162.18

Was this useful?

Codacy

3 releases · seen 2026-08-08 Notes

Codacy is an automated code analysis platform that identifies bugs, security vulnerabilities, and code quality issues across repositories.

Codacy added two new analysis engines — GolangCI-Lint for Go and PHP CS Fixer for PHP code style — and extended repository config-file support to Checkov and ShellCheck so their settings can be version-controlled instead of managed only through the UI.

└──▷ WHAT SHIPPED · 3 FEATURES 0 OF 3 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Repository config file support for Checkov and ShellCheck new 81

Codacy can now read analysis settings directly from repository config files instead of requiring UI-only configuration: enabling "Use Checkov configuration file found in the repository" under Repository > Code Patterns > Checkov picks up a .checkov.yaml or .checkov.yml file, and enabling "Use Shellcheck configuration file found in the repository" under Repository > Code Patterns > Shellcheck picks up a .shellcheckrc file in the repository root.

Let Codacy pick up your existing Checkov config instead of managing every check through the UI.

📍Repository › Code Patterns › Checkov › enable "Use Checkov configuration file found in the repository"

Commit a .shellcheckrc to your repo and enable file-based analysis so ShellCheck rules are version-controlled alongside your code.

📍Repository › Code Patterns › Shellcheck › enable "Use Shellcheck configuration file found in the repository"

— Exact file names and UI toggle paths for both engines. Adding PHP CS Fixer and support to Checkov c…Support for Shellcheck configuration file - …

02 00

PHP CS Fixer engine for PHP code style new 78

Codacy adds PHP CS Fixer as a new supported engine for PHP code-style analysis, configurable via the Codacy UI or via your repository's own .php-cs-fixer.dist.php or .php-cs-fixer.php file. When both exist, Codacy-selected patterns in the UI take precedence over the repository config file.

— Names exact config files and the precedence rule between them. Adding PHP CS Fixer and support to Checkov c…

03 00

GolangCI-Lint support for Go analysis new 60

Codacy adds GolangCI-Lint as a client-side analysis tool for Go projects: run it locally or in CI and upload results to the Codacy dashboard for reporting. It is not enabled by default on existing repositories and must be turned on manually in repository settings to avoid a large volume of new issues appearing unexpectedly.

— Mentions repository settings but no exact path or command. Adding GolangCI-Lint as new supported tool J…

Was this useful?

StackHawk

1 release · seen 2026-08-08 Notes

StackHawk is an application security testing platform that performs dynamic security scanning to find vulnerabilities in web applications during development and CI/CD pipelines.

StackHawk's HawkScan 6.3.0 release adds mid-scan crash detection and a much faster replay-based rescan, alongside new seedPaths validation and a consolidated CLI credential model, while the platform API also changed the audit endpoint's response schema.

└──▷ WHAT SHIPPED · 5 FEATURES 3 OF 5 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Crash detection for scans new 91

A new hawk.scan.crashDetection config block detects when a target application crashes or becomes unresponsive mid-scan by passively counting new timeouts and connection failures, then triggering an active liveness probe (using app.waitForAppTarget as the probe path) before taking action. The trigger threshold is set via hawk.scan.crashDetection.maxTimeouts, which defaults to 5x concurrentRequests when unset or zero, and the response is controlled by hawk.scan.crashDetection.action, supporting ABORT (default, fails the scan with error scan.target-crashed), WARN (logs and continues), or DISABLED.

Abort a scan immediately when the target goes down, using a tighter failure threshold than the default to catch crashes faster.

yaml

hawk:
  scan:
    crashDetection:
      maxTimeouts: 10
      action: ABORT

Abort a scan immediately and surface a clear error when the target application stops responding mid-scan, rather than burning the full scan window against a dead host.

yaml

hawk:
  scan:
    crashDetection:
      action: ABORT
      maxTimeouts: 10

— Full mechanism, exact config keys, defaults, and runnable YAML example. snapshot-20260808

02 01

Faster rescans via request replay breaking 79

hawk rescan now replays the exact requests that produced findings — fetched from the platform — instead of re-crawling the app, skipping all discovery (OpenAPI, SOAP, GraphQL, HAR imports) for substantially faster, tightly scoped retests; it falls back to a normal spider automatically if replay cannot run. Because discovery is now skipped entirely, workflows that relied on rescan to re-crawl the application must switch to a full scan instead.

— Clear before/after mechanism and named command, no runnable example given. snapshot-20260808

03 00

seedPaths validation at parse time new 78

hawk.spider.seedPaths entries are now validated at configuration parse time by both hawk scan and hawk validate config: entries over 200 characters or with more than 12 path segments produce a warning, while entries over 1000 characters fail validation outright.

— Exact config key and thresholds named, but no example run shown. snapshot-20260808

04 02

Unified CLI credential sourcing breaking 75

Commands embedded in the HawkScan CLI (hawk op) now read API key, API host, and organization from the same sources as hawk itself — environment variables and ~/.hawk/hawk.properties — eliminating the need for a separate ~/.hawkop/config.yaml, which is no longer used.

— Names exact config file paths and the removed file; clear migration note. snapshot-20260808

thinner coverage below

05 03

Audit endpoint response schema changed breaking 33

The GET /api/v1/org/{orgId}/audit endpoint had its response schema changed.

— Endpoint named but no detail on what changed in the schema. 0.0.1

└──▷ BREAKING ON UPGRADE

!
hawk rescan no longer re-runs discovery — all discovery is skipped, including OpenAPI, SOAP, GraphQL, and HAR imports. Workflows that relied on rescan to re-crawl the application will need to use a full scan instead.
!
The separate ~/.hawkop/config.yaml credentials file is no longer used; credentials must now be in environment variables or ~/.hawk/hawk.properties.
!
The separate ~/.hawkop/config.yaml credential file is no longer used by hawk op; credentials are now read exclusively from the environment and ~/.hawk/hawk.properties.

Was this useful?

Vercel deepsec

changes since 2026-07-09 Code

Deepsec is a security harness for finding vulnerabilities in your codebase powered by coding agents

deepsec gains --thinking-level, one-shot initialization, stable finding IDs, and a broadened SSRF matcher.

Adds --thinking-level CLI flag to control agent reasoning depth during scans.
Ships one-shot Deepsec initialization (PR #139) to bootstrap a workspace in a single command.
Introduces stable finding IDs and permissive verdict reconciliation in the revalidate workflow, backed by ensureFindingIds and reconcileVerdicts.
Broadens the SSRF matcher to cover a wider range of server-side request forgery patterns.

4 details in Everything

Was this useful?

ElcomSoft

1 release · 2026-08-08 Changelog

ElcomSoft develops digital forensics, password recovery, and mobile device extraction tools for law enforcement and security professionals.

ElcomSoft shipped EIFT Firewall, a free macOS companion app that replaces its 2023 firewall script for safely sideloading the iOS Forensic Toolkit's low-level extraction agent.

└──▷ WHAT SHIPPED · 1 FEATURE 1 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

EIFT Firewall macOS companion app new 58

New free macOS application EIFT Firewall creates a temporary, controlled USB network path between a Mac and a connected evidentiary device to safely sideload the Elcomsoft iOS Forensic Toolkit low-level extraction agent, replacing the macOS firewall script published in 2023.

— Names mechanism and predecessor but no install/usage steps. changelog-20260808-e8228529

Was this useful?

quark-engine

1 release · 2026-08-08 Notes

Quark-Engine is a static analysis tool for Android malware detection that identifies suspicious behaviors and potential security threats in APK files.

Quark-Engine v26.8.1 adds AI-assisted analysis skills for Claude Code, a new DexTrace backend for DEX analysis, and fresh malware analysis reports.

└──▷ WHAT SHIPPED · 3 FEATURES 0 OF 3 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Claude Code skills for malware analysis and rule generation new 60

Adds Claude Code skills /quark and /quark-rule-gen, enabling AI-assisted malware analysis and rule generation directly within Claude Code.

— Names two slash commands but no mechanism detail. v26.8.1

thinner coverage below

02 00

Malware analysis reports for AhRat, AndroRAT, and Sova new 40

Adds Quark malware analysis reports for AhRat, AndroRAT, and Sova.

— Names three malware families but no report contents or access path. v26.8.1

03 00

DexTrace core-library backend for DEX analysis new 25

Adds a new DexTrace core-library backend for DEX analysis.

— Bare name, no mechanism or usage given. v26.8.1

Was this useful?

CAPEv2

changes since 2026-07-20 Code

Malware Configuration And Payload Extraction

CAPEv2 adds Remus malware detection & config extraction, a ThreatFox/Malpedia threat-intel module, and PCI vendor ID spoofing to defeat QEMU fingerprinting

New modules/processing/threatintelligence.py processing module (runs at order=22) enriches analysis reports with two phases: an indicator phase that queries ThreatFox to annotate contacted IPs/domains with threat tags, and a family phase that fetches descriptive cards (name, aliases, description, references, attribution) from Malpedia for detected malware families; results are written to results["threatintelligence"] and surfaced in a collapsible Threat Intelligence report-page section.
Adds Remus malware detection and dynamic config extraction.
Adds PCI vendor ID spoofing to bypass Al-Khaser QEMU detection checks.
New require_task_delete / can_delete_task authorization guards in the multi-tenant API enforce that only the submitter or a box-admin can delete a task (public policy).
The API v2 batch-delete endpoint now reports partial failures per task (individual try/except per task in batch loops) and returns a partial_error field instead of a bodiless 500, with SQL delete committed before irreversible folder/Mongo deletes for rollback safety.

5 details in Everything

└──▷ BREAKING ON UPGRADE

!The suspended parameter has been removed from execute() methods in process.py and abstracts.py; callers passing this argument will break.

Was this useful?

Radare2

1 release · 2026-08-07 Notes Code

UNIX-like reverse engineering framework and command-line toolset

Radare2 6.2.0 ships a new generic binary resource extraction API alongside broader calling-convention and register analysis for ppc64 ELFv2 and MIPS, an initial Rust v0 demangler, WebAssembly type support, and several breaking changes including a Capstone v4 minimum and renamed resource commands.

└──▷ WHAT SHIPPED · 12 FEATURES 3 OF 12 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

Binary resource extraction and generic resource API breaking 95

Adds iuo to open binary resources inline and ix to extract sections, segments, and resources; adds iU (with iUx extraction) and rabin2 -xU for listing/extracting resources. The old iR command is renamed to iU, and the previous iU command moves to iu; the rabin2 -u flag behavior moves to an environment variable. A new generic binary resource API adds support for GLib Resources in ELF, Macho code signatures, PE/CLR resources, PE resources from MDMP, OS/2 LE/LX resources, QNX LMF resources, and Pebble resources.

Extract all embedded resources from a PE binary directly from the command line.

$ rabin2 -xU malware.exe

— Names every command, flag, and rename with runnable example. 6.2.0

02 00

Calling convention and register analysis improvements new 85

Supports the ppc64 ELFv2 calling convention, selected automatically from the ELF ABI, and selects the MIPS calling convention automatically from the ELF o32/n32 ABI. Adds callconv definitions (clobber and preserve registers) for all supported architectures including PPC and ARM, PPC64 TOC pointer base-register tracking gated by RArchOpMask.STATEFUL and wired into aar, new r_anal_cc_argslot, r_anal_cc_argval, and r_anal_cc_isclobber APIs, RArchOpMask.STATEFUL and RArchSession.reset for stateful architecture decoding in plugins, and reguse analysis hints with dyncc register role APIs.

— Lists APIs and mechanisms but no direct usage command. 6.2.0

03 00

New configuration keys for analysis and UI new 80

Adds anal.gpseed to seed a PIC base register during ESIL analysis, asm.cmt.reguse to emit register-use comments in disassembly, dir.binsdb for PE/MDMP ordinal lookups from binpath/dirprefix, types.fields to gate propagation of callee types into struct and union fields during aaft, and scr.drag to enable click-and-drag scrolling in disasm, hex, and graph views.

Seed a PIC base register before ESIL-based analysis so position-independent code resolves correctly.

$ r2 /bin/firmware.elf
> e anal.gpseed=0x10000
> aaa

— Names every key and its file/effect; one has runnable example. 6.2.0

04 00

Struct type synthesis via afts command new 80

Adds struct type synthesis from pointer-argument access patterns via the afts command, letting analysts recover unknown data structures after analyzing a function.

Synthesize struct types from pointer-argument access patterns after analyzing a function to recover unknown data structures.

$ r2 /bin/target
> aaa
> s sym.process_packet
> afts

— Named command with runnable example demonstrating use. 6.2.0

05 02

r2agent HTTP auth protection breaking 60

Adds auth protection to r2agent web session spawns and protects r2agent with an HTTP auth file; empty lines in the auth file now block all access rather than allowing any password, reversing prior behavior.

— Names the auth file mechanism and behavior change. 6.2.0

thinner coverage below

06 00

Short-circuit && and || chaining in r2 shell new 50

Adds && and || short-circuit command chaining support in the r2 shell, including after literal commands in the shell pipeline.

— Names the operators but no worked example. 6.2.0

07 00

Rust v0 demangler and pluggable demangler framework new 45

Adds an initial Rust v0 symbol demangler, making Rust binaries easier to analyze, and makes RBin demanglers pluggable and selectable.

— Describes capability but no command shown. 6.2.0

08 00

Lcc command to list registered plugins new 40

Adds Lcc command to list all plugins registered with the new command API.

— Named command, minimal further detail. 6.2.0

09 03

Capstone v4 required, v3 removed breaking 40

Removes Capstone v3 support; Capstone v4 is now the minimum required version for disassembly.

— Clear before/after but no migration steps. 6.2.0

10 00

New build targets: MiKNix and jslinux new 35

Adds MiKNix platform support in the build system and a dist/jslinux target using temu/rv64 for JavaScript-based Linux builds.

— Names targets but no build instructions. 6.2.0

11 00

RBinPluginAny renamed and LibMagic dropped breaking 35

RBinPluginAny is renamed to "null" and no longer uses LibMagic.

— Names the rename but no impact guidance. 6.2.0

12 00

WebAssembly v128, ref and reftype support new 30

Adds v128, ref, and reftype support for WebAssembly binaries.

— Names the types but no mechanism or example. 6.2.0

└──▷ BREAKING ON UPGRADE

!
The iR command is renamed to iU; the old iU command is moved to iu; the rabin2 -u flag behavior is moved to an environment variable.
!
Capstone v3 support is removed; Capstone v4 is now the minimum supported version.
!
RBinPluginAny is renamed to "null" and no longer uses LibMagic.
!
Empty lines in the r2agent HTTP auth file now deny all access (previously they permitted any password through).

Was this useful?

Fastly

1 release · 2026-08-07 Blog

Fastly is a content delivery network and edge computing platform that accelerates website performance and provides security services globally.

Fastly CLI v15.6.0 adds a new command group for managing audit log event mappings.

└──▷ WHAT SHIPPED · 1 FEATURE 0 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Audit log event-mapping command group in CLI new 60

The Fastly CLI adds the audit-log event-mapping command group for managing audit log event mappings.

— Names exact command but not its subcommands or behavior. CLI v15.6.0 (2026-08-07)

Was this useful?

Plerion

1 release · seen 2026-08-08 Notes

Plerion is a cloud security platform that enables continuous vulnerability discovery and remediation across cloud infrastructure and applications.

Plerion significantly deepened its cloud security posture coverage for Google Cloud AI services, adding dozens of new checks, guardrails, and expanded asset discovery across the GCP AI stack.

└──▷ WHAT SHIPPED · 1 FEATURE 1 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

Expanded GCP AI security posture coverage new 63

Plerion adds 46 new posture checks spanning Vertex AI, Vertex AI Workbench, Colab Enterprise, Agent Engine, Vertex AI Search, Gemini Enterprise, Model Armor, Dialogflow CX, Cloud TPU, and Gemini Code Assist. It also adds project-level org-policy guardrails covering Vertex AI Workbench access and Model Garden model use, and extends asset discovery to 16 new GCP resource types behind these newly supported AI services.

— Names all services and checks but no config or command to act on. GCP AI coverage widens further:

Was this useful?

NVIDIA OpenShell

1 release · 2026-08-07 Notes

OpenShell is the safe, private runtime for autonomous AI agents.

OpenShell v0.0.101 focused on build system plumbing, adding Bazel build targets for two internal drivers.

└──▷ WHAT SHIPPED · 1 FEATURE 0 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Bazel build targets for credential and VM drivers new 35

Adds Bazel build targets for the credential driver, enabling it to be compiled as part of the standard build, and adds a VM driver built from GitHub sources via Bazel.

— Names build system only; no usage or config detail given. v0.0.101

Was this useful?

Sysdig Secure

1 release Blog

Sysdig Secure is a container security platform providing runtime monitoring, vulnerability scanning, and compliance enforcement for cloud-native applications.

Sysdig Secure improved vulnerability policy scoping, allowing the same attribute to be used in multiple AND-joined conditions for Admission Control and Runtime stages.

└──▷ WHAT SHIPPED · 1 FEATURE 0 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Multi-condition attribute scoping in vulnerability policies improved 55

Vulnerability policy scopes for the Admission Control and Runtime stages now support using the same attribute in more than one condition, joined by AND, enabling include and exclude patterns on a single attribute (e.g., all non-production clusters except one).

— Explains mechanism and gives example but no config path Secure SaaS - August 3, 2026

Was this useful?

Google SecOps

1 release · 2026-08-08 Changelog

Google SecOps is a unified security operations platform combining threat detection, investigation, and response capabilities across cloud and enterprise environments.

Google SecOps expanded its log ingestion catalog with 25 new supported sources spanning security, network, and SaaS vendors, and added self-service license key retrieval for Bindplane Enterprise collection agents.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Self-service Bindplane Enterprise license key download new 75

Enterprise Plus and Google Unified Security (GUS) customers can now download the Bindplane Enterprise (Google Edition) license key directly from SIEM Settings > Collection Agents in the console, without opening a support ticket. This capability is in Preview.

Retrieve your Bindplane Enterprise license key from the console so you can activate a self-managed collection agent without opening a support ticket.

📍In the Google SecOps console, go to SIEM Settings › Collection Agents and download the Bindplane Enterprise (Google Edition) license key.

— Names exact UI path and preview status, gives steps. changelog-20260808-95ed47a5

02 00

26 new log ingestion source integrations new 60

Google SecOps added support for ingesting logs from a broad set of new sources: ARMIS_ALERTS (Armis Alerts), GCP_SECURITYCENTER_EXTERNAL_EXPOSURE (Security Command Center External Exposure), ADOBE_EXPERIENCE_PLATFORM (Adobe Experience Platform), AUDIOCODES_SBC (AudioCodes Session Border Controller), AZURE_GATEWAY_CONTAINERS (Azure Application Gateway for Containers), AZURE_LOGIC_APPS (Azure Logic Apps), AZURE_NATGW_FLOW (Azure NAT Gateway Flow), BROADCOM_DX_NETOPS_SPECTRUM (Broadcom DX NetOps Spectrum), CARTO_ACTIVITY (Carto Activity), CLAUDE_CODE_OBSERVABILITY (Claude Code Observability), CYBLE_ASM (Cyble Attack Surface Management), CYBLE_BIP (Cyble Brand Intelligence & Protection), DARKWEB_IQ (Darkweb IQ), ELLIO_THREAT_INTEL (Ellio Threat Intelligence), EXEON_NDR (Exeon NDR), GRAVITEE (Gravitee), KASPERSKY_ANTI_TARGETED_ATTACK (Kaspersky Anti Targeted Attack), MICROSOFT_COPILOT_INTERACTION (Microsoft Copilot Interaction), OSTTRA_MARKITWIRE (OSTTRA MarkitWire), PROOFPOINT_ADAPTIVE_EMAIL_SECURITY (Proofpoint Adaptive Email Security), SECOMEA_GATEMANAGER (Secomea GateManager), TRENDMICRO_VISION_ONE_RISK_EVENT (Trend Micro Vision One Risk Event), TXONE_EDGEIPS (TXOne EdgeIPS), VECTRA_RUX (Vectra Respond UX), and ZOHO_CRM (Zoho CRM).

— Names every source ID but no ingestion mechanism detail. changelog-20260808-95ed47a5

└──▷ ALSO FROM THESE RELEASES

Was this useful?

Elastic Security Labs

1 release · 2026-08-07 Blog

Elastic Security Labs publishes security research, threat analysis, and vulnerability findings to help organizations understand emerging cyber threats.

Elastic Security Labs shipped a purpose-built Elastic Agent integration that watches for silent removal of npm's cooldown protection (min-release-age) by snapshotting .npmrc files across hosts every six hours.

└──▷ WHAT SHIPPED · 1 FEATURE 1 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

CEL integration detecting npm cooldown removal via .npmrc snapshots new 90

A roughly 40-line Common Expression Language (CEL) integration in Elastic Agent snapshots every .npmrc on a 6-hour heartbeat and sets cooldown.absent = true in the ingest pipeline when the min-release-age setting disappears between snapshots. It uses the file() and dir() CEL functions to enumerate user home directories (/Users, /home) and machine-global npm config paths (/opt/homebrew/etc/npmrc, /usr/local/etc/npmrc, /etc/npmrc, /usr/lib/node_modules/npm/.npmrc) covering both macOS and Linux from one script, with global paths and poll interval supplied via the Custom request cursor YAML field in the Fleet UI, and resource.url: file:///dev/null used to satisfy the integration's required URL field without issuing an HTTP request. Registry auth token lines (//registry.npmjs.org/:_authToken=...) are filtered out in the ingest pipeline before leaving the host, and per-file file.hash.sha256 values are emitted alongside file content so in-place rewrites that strip min-release-age without changing the file path can still be detected. The write-up also documents why the Custom Logs Filestream integration was unsuitable: files under 1024 bytes are silently held back under the default fingerprint identity strategy, filestream only emits on append (not on content removal), and its clean_inactive option requires ignore_older to also be set, which suits log rotation but not static config files that may sit unchanged for months.

— Names exact paths, YAML field, pipeline keys and fallback rationale. The security signal log tailing can't see: t…

Was this useful?

Iru

1 release · seen 2026-08-08 Notes

Iru's August 7th update adds a new library item assignment criterion based on Apple beta program enrollment and refreshes the Iru AI chat experience.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Apple beta enrollment as assignment criterion new 40

Library items can now be assigned based on a device's Apple beta program enrollment status, adding this as a new criterion for library item assignment.

— Names the criterion but no config path or mechanism given. What's New in Iru | August 7th

02 00

Refreshed Iru AI chat experience improved 20

Iru's AI chat experience has been refreshed to provide faster, smarter responses.

— Vague marketing phrasing with no specifics on mechanism or scope. What's New in Iru | August 7th

Was this useful?

Huntress

1 release · seen 2026-08-08 Notes

Huntress is a managed security platform that provides threat detection, incident response, and endpoint protection for small and mid-sized businesses.

Huntress rolled out a redesigned Managed ITDR Dashboard in Early Access, adding rapid identity triage, failed login characterization, and a built-in SIEM search surface.

└──▷ WHAT SHIPPED · 4 FEATURES 3 OF 4 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

Quick SIEM Search in ITDR dashboard new 73

A new Quick SIEM Search surface within the ITDR dashboard provides pre-populated queries against the full Entra Unified Audit Log, stored in Huntress SIEM at no cost for ITDR customers for up to one year.

— Names retention and cost terms but no query syntax given Now Live: Managed ITDR Dashboard in Early Ac…

02 02

Rapid Identity Triage panel new 72

A new panel lets an analyst enter a user's email to get an immediate overview of all activity over the past 24 hours, with options to export a timeline, revoke sessions, or disable the account.

— Names concrete actions but no exact UI path or command Now Live: Managed ITDR Dashboard in Early Ac…

thinner coverage below

03 03

Failed Login Characterization view new 50

A new view shows failed login attempts globally, with categorization by residential proxies, VPNs, and datacenters.

— Describes categories shown but no interaction detail Now Live: Managed ITDR Dashboard in Early Ac…

04 00

Redesigned ITDR dashboard navigation new 50

The redesigned ITDR dashboard is accessible via the ITDR icon in the left-hand navigation, and the existing dashboard remains available in parallel during Early Access.

— Gives exact navigation path but no further mechanism Now Live: Managed ITDR Dashboard in Early Ac…

Was this useful?

Trend Vision One

2 releases · 2026-08-04 → 2026-08-06 Blog

Trend Vision One is Trend Micro's unified cybersecurity platform providing threat intelligence, vulnerability management, and incident response across endpoints and networks.

Trend Vision One's Agentless Vulnerability and Threat Detection expands scan coverage to Azure Function Apps and adds customer-configurable resource tagging for OCI deployments.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Customer-configured OCI resource tagging via Terraform new 65

Agentless Vulnerability & Threat Detection v1.96.3a0922b adds support for customer-configured tags — both OCI defined tags and freeform tags — applied to all Terraform-deployed OCI resources, including the scanner VM, disk snapshots, and cloned scan disks.

— Names resource scope but no exact config key or command. Agentless Vulnerability & Threat Detection v…

thinner coverage below

02 00

Azure Function Apps scanning in AVTD new 55

Agentless Vulnerability and Threat Detection now scans Azure Function Apps for vulnerabilities and malware, accessible via Cyber Risk Exposure Management → Continuous Risk Management → Threat and Exposure Management.

— Names the UI path but no mechanism or limits given. Agentless Vulnerability and Threat Detection…

Was this useful?

OpenCTI

1 release · 2026-08-07 Notes

Open Cyber Threat Intelligence Platform

OpenCTI 7.260807.0 introduces shareable saved searches and dashboard filters, direct XTMHub connectivity, advanced Lucene search operators, and several UI and ingestion improvements.

└──▷ WHAT SHIPPED · 6 FEATURES 0 OF 6 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Lucene wildcard and fuzzy search operators new 60

Adds support for Lucene advanced search syntax including wildcard prefix (*) and fuzzy search (~) operators.

— Names exact operators reader can use directly in search. 7.260807.0

thinner coverage below

02 00

Admin UI for custom fields configuration new 45

Adds an Admin UI (Settings) for custom fields configuration, described as Chunk 2 of the custom-fields feature.

— Names the settings location but lacks specific fields or flags. 7.260807.0

03 00

Shareable saved searches and reusable filters new 35

Adds shareable and reusable saved searches across the platform, plus reusable saved filters inside dashboards and custom views.

— Describes what's shared but no UI path or config detail. 7.260807.0

04 00

XTMHub connectivity from OpenCTI interface new 35

Adds XTMHub connectivity, allowing users to connect to Hub directly from the OpenCTI interface.

— Names the integration but not how it's configured. 7.260807.0

05 00

Label casing preserved in UI improved 35

Preserves user-defined label casing, removing the automatic text-transform: capitalize styling previously applied to labels.

— Names the CSS rule removed but minor cosmetic change. 7.260807.0

06 00

TAXII feed tabs in ingestion UI new 30

Adds TAXII feed tabs to the ingestion UI for improved feed management.

— Bare description of a UI addition without mechanism. 7.260807.0

Was this useful?

Tenzir

1 release · 2026-08-07 Blog

Tenzir is a security data platform that ingests, transforms, and analyzes security telemetry and threat intelligence data.

Tenzir Platform v1.38.0 makes telemetry collection opt-in and lets users search pipelines by ID.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Pipeline search by ID improved 50

The pipeline list search box now matches on pipeline ID in addition to name and definition, letting you paste a pipeline ID to jump straight to it.

— Names the UI mechanism but no exact navigation path Tenzir Platform v1.38.0: Opt-in telemetry co…

02 00

Opt-in telemetry consent on Account page breaking 50

Telemetry collection is now gated behind explicit user opt-in, configurable through the Account page, and is now also enabled on app.tenzir.com for opted-in users.

— Names the Account page setting but no exact steps Tenzir Platform v1.38.0: Opt-in telemetry co…

Was this useful?

Twingate

1 release · 2026-08-08 Changelog

Twingate is a zero-trust network access platform that securely connects remote users and devices to private resources without VPN.

Twingate's Admin Console now lets admins resize table columns, with sizing preferences saved per user.

└──▷ WHAT SHIPPED · 1 FEATURE 0 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Resizable table columns in Admin Console improved 55

Admins can resize table column widths in the Admin Console by hovering over a column edge and dragging. Column sizing is saved per-user, and a reset-to-default option is available.

— Names UI mechanism and behavior but no config/API surface changelog-20260808-d9a735cb

Was this useful?

Cohesive VNS3 Controller

1 release · seen 2026-08-08 Notes

VNS3 Controller is a network security appliance that creates encrypted virtual networks and manages multi-cloud connectivity with advanced firewall capabilities.

VNS3 7.2.1 upgrades the underlying kernel to address a privilege escalation issue and adds platform-awareness to the Plugin Catalog.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Kernel upgraded to Linux 7.0.0 improved 50

VNS3 upgraded from the Linux 6.17.x kernel line to 7.0.0 in response to reported local privilege escalation findings against the 6.17 line.

— Names versions and reason but no reader action snapshot-20260808

02 00

Plugin Catalog platform and version awareness improved 40

The Plugin Catalog now differentiates between ARM and Intel capable plugins, reflecting potential release timeline differences between platforms, and supports more intelligent version management with listings enhanced to show multiple versions and platform compatibility information.

— Describes what changed but no UI path or exact fields snapshot-20260808

Was this useful?

Sophos ZTNA

2 releases · seen 2026-08-08 Notes

Sophos ZTNA is a zero-trust network access solution that verifies user and device identity before granting access to applications and resources.

Sophos ZTNA 2025.02 added Let's Encrypt certificate automation and expanded to all Sophos Central regions, while 2025.23 improved on-premises connectivity with traffic bypass and domain controller support.

└──▷ WHAT SHIPPED · 4 FEATURES 1 OF 4 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

Let's Encrypt certificate generation via Sophos Central new 55

ZTNA gateways can now generate Let's Encrypt certificates directly through Sophos Central, with automatic management and renewal handled by Sophos.

— Names mechanism and scope but no config steps 2025.02

02 00

On-premises traffic bypass for the ZTNA agent new 55

A new configuration option prevents the ZTNA agent from intercepting on-premises traffic, avoiding tunnel hairpinning for local resources.

— Explains behaviour and benefit but no exact setting name 2025.23

03 00

Domain controller as agent-based resource new 45

ZTNA now supports adding a domain controller as an agent-based resource, providing seamless connectivity to resources behind it.

— Describes new resource type without configuration steps 2025.23

04 00

ZTNA expanded to all Sophos Central regions improved 25

Sophos ZTNA availability is expanded to all Sophos Central regions.

— Bare statement of expanded availability, no detail 2025.02

Was this useful?

Auth0

1 release · 2026-08-07 Blog

Auth0 is a cloud-based identity and access management platform providing authentication, authorization, and user management solutions for applications.

Auth0 shipped two identity-management additions this window: new My Account API endpoints for managing a user's own authentication methods, and an Early Access Session Delegation capability built on Custom Token Exchange.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Authentication methods endpoints in My Account API new 90

The My Account API gains four endpoints for managing a user's enrolled authentication methods: GET https://{yourDomain}/me/v1/authentication-methods to list them, POST https://{yourDomain}/me/v1/authentication-methods to create/enroll a new one, POST https://{yourDomain}/me/v1/authentication-methods/{id}/verify to verify a specific method by ID, and DELETE https://{yourDomain}/me/v1/authentication-methods/{id} to remove one.

List all authentication methods enrolled for the currently authenticated user.

$ curl -X GET 'https://{yourDomain}/me/v1/authentication-methods' \
  -H 'Authorization: Bearer <access_token>'

Verify a specific authentication method (e.g., after enrollment) to confirm it is active.

$ curl -X POST 'https://{yourDomain}/me/v1/authentication-methods/{id}/verify' \
  -H 'Authorization: Bearer <access_token>'

— Four full endpoints with methods and paths, plus runnable examples. product docs

02 00

Session Delegation via Custom Token Exchange new 81

In Open Early Access, requesting a Custom Token Exchange token with a session_transfer audience returns a Session Transfer Token (instead of an access token), authorized by the same Action logic already controlling delegation; delegated sessions are redeemed through the standard /authorize endpoint with no extra login step for the target user. Both identities are preserved via the sub claim (the session owner) and the act claim per RFC 8693 (the acting party, e.g. a support agent), and dedicated tenant log events record delegated logins separately from regular logins for compliance and traceability.

— Mechanism and claims fully explained but no runnable example given. Custom Token Exchange - Session Delegation i…

Was this useful?

OpenAI Codex CLI

2 releases · 2026-08-07 → 2026-08-08 Notes Code

Lightweight coding agent that runs in your terminal

OpenAI Codex CLI's latest alpha releases add session forking and a migrate-rollouts command for thread history, Markdown conversation export and archiving in the TUI, a Guardian circuit-breaker for cyber-specialty models, and expanded MCP hook/event support alongside agent-identity endpoint overrides and several smaller diagnostics and model-list additions.

└──▷ WHAT SHIPPED · 0 FEATURES13 FEATURES 0 OF 13 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Session forking in codex exec new 93

Adds a fork subcommand to codex exec that branches a previous session into a new one, accepting a SESSION_ID, optional --image/-i file attachments, and an optional PROMPT (or - to read from stdin).

Branch off a previous session to explore an alternative approach without losing the original thread.

$ codex exec fork session-123 "try a different refactor approach"

Fork a session and attach a screenshot as context for the follow-up prompt.

$ codex exec fork session-123 --image screenshot.png "fix the layout issue shown in the image"

— Full flag list plus two runnable command examples. rust-v0.148.0-alpha.5 rust-v0.148.0-alpha.2

02 00

Agent identity endpoint overrides new 80

Honors CODEX_AGENT_IDENTITY_AUTHAPI_BASE_URL and CODEX_AGENT_IDENTITY_JWKS_BASE_URL environment variables to override agent identity registration and verification endpoints.

— Names exact environment variables to set. rust-v0.148.0-alpha.2

03 00

MCP tool hook handler configuration new 78

Recognizes mcp_tool hook handler configurations (with server, tool, JSON input, and optional timeout and status message) and returns them from configRequirements/read.

— Names config fields and the read endpoint. rust-v0.148.0-alpha.2

04 00

Markdown conversation export in TUI new 75

Adds a /export command in the TUI that exports the full conversation history as structured Markdown, supporting clipboard and file destinations, default filename prompting, and an explicit path argument.

— Names command and destinations but no worked example. rust-v0.148.0-alpha.5 rust-v0.148.0-alpha.2

05 00

wait_agent timeout clamping improved 70

Clamps wait_agent timeout requests below min_wait_timeout_ms to the configured minimum instead of rejecting them, and reports the adjustment in the tool output.

— Named parameters but no user-facing command. rust-v0.148.0-alpha.2

thinner coverage below

06 00

Session archiving and restore in resume picker new 40

Adds session archiving to the resume picker in the TUI, allowing sessions to be archived and restored.

— Describes behaviour but no command or flag named. rust-v0.148.0-alpha.5 rust-v0.148.0-alpha.2

07 00

Diagnostics: app-server API and process snapshots new 35

Exposes app-server diagnostics through the experimental API and adds process diagnostics snapshots.

— Groups two thin diagnostics additions without endpoint detail. rust-v0.148.0-alpha.5

08 00

Multi-agent versions in model/list improved 35

Exposes multi-agent versions in model/list.

— Names endpoint field but no further detail. rust-v0.148.0-alpha.5

09 00

Call IDs in MCP requests improved 21

Includes call IDs in MCP requests.

— Bare statement without field name or usage. rust-v0.148.0-alpha.5

10 00

MCP event discovery and subscriptions new 18

Adds MCP event discovery and subscriptions.

— Bare description with no mechanism or endpoint. rust-v0.148.0-alpha.5

11 00

Resilient response streams on connection failure improved 18

Keeps response streams alive through connection failures.

— States behaviour without mechanism or scope. rust-v0.148.0-alpha.5

12 00

Unified image budget across agents new 16

Adds a unified image budget across agents.

— No limit, mechanism, or config named. rust-v0.148.0-alpha.5

13 00

Default code-mode exec yield timeout improved 16

Configures the default code-mode exec yield timeout.

— No value or config key given. rust-v0.148.0-alpha.5

Was this useful?

Claude Code

1 release · 2026-08-08 Notes Code

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

v2.1.225 improves gateway spend-limit warnings and expands Remote Control session messaging, alongside a new workspace trust prompt for claude agents.

└──▷ WHAT SHIPPED · 3 FEATURES 0 OF 3 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Remote Control session messaging improvements improved 70

SendMessage can now initiate a conversation with Remote Control sessions on other machines by name, with ListAgents surfacing them as name [ref], instead of only being able to reply after they message you first. SendMessage also now locks in a Remote Control recipient you have already confirmed, preventing it from being silently swapped for a same-named session on the local machine when the remote list could not be checked. Separately, photos attached from the Claude app are now passed directly to Claude rather than being read from disk via a separate tool call.

— Names APIs and format string, explains before/after behavior v2.1.225

thinner coverage below

02 00

Gateway spend-limit warnings improved 55

Claude Code's usage warning now supports gateway spend limits — when the limit is reached, the message names the cap, its reset time, and the operator's message. This requires the gateway to be on version 2.1.225.

— Names cap/reset/message fields but no config surface shown v2.1.225

03 00

Workspace trust prompt in claude agents improved 40

claude agents now shows a workspace trust prompt for untrusted directories, matching the existing behavior of claude.

— Names the command but no detail on prompt options v2.1.225

Was this useful?

OpenHands

1 release · 2026-08-07 Notes

OpenHands: AI-Driven Development

OpenHands v1.11.0 adds per-run LLM cost visibility, automation tag filtering, and a typed action for launching child conversations, alongside several UI refinements.

└──▷ WHAT SHIPPED · 5 FEATURES 0 OF 5 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Typed action for child conversations new 55

Adds a typed agent action for launching local or Cloud child conversations, enabling programmatic multi-conversation workflows.

— Explains purpose and scope but no exact API name. v1.11.0

02 00

Per-run LLM cost visibility new 50

Shows per-run LLM cost in the Activity Log and in exports, giving teams immediate spend visibility for each agent run.

— Names Activity Log and exports but no metrics or limits. v1.11.0

03 00

Tag metadata in conversations list improved 45

Adds tag chips, overflow handling, and hovercard labels to the conversations list for richer conversation metadata display.

— Names three specific UI elements added to the list. v1.11.0

04 00

UI polish across Automations, Customize, and Agent Canvas improved 40

Keeps the Automations pane visible even when empty to reduce navigation confusion, reorders the Customize navigation section for improved discoverability, and polishes the Agent Canvas version-update UI in Settings.

— Names three UI areas but each change described thinly. v1.11.0

05 00

Automation tag filtering and recognition new 35

Adds an automation tag filter and recognition so automations can be organized and surfaced by tag.

— States the capability but no mechanism or UI path detail. v1.11.0

Was this useful?

LangGraph

1 release · 2026-08-07 Notes

Build resilient agents.

LangGraph's checkpoint package gains an opt-in flag to filter out expired state rows when reading checkpoints.

└──▷ WHAT SHIPPED · 1 FEATURE 0 OF 1 FEATURE most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Opt-in omit_expired flag for checkpoint reads new 70

The omit_expired parameter can be passed to checkpoint reads, supported in the checkpoint and checkpoint-postgres packages, to skip expired rows and prevent stale state from surfacing in long-running workflows.

— Names exact flag and packages, but no usage example. checkpoint==4.2.0

Was this useful?

Daytona

1 release · seen 2026-08-08 Notes

Daytona is a development environment platform that enables standardized, reproducible coding spaces for teams and AI agents.

Daytona 0.199.0 adds time-to-live control for sandboxes and raises the minimum supported Python version for its SDK.

└──▷ WHAT SHIPPED · 2 FEATURES 0 OF 2 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Sandbox TTL support in SDK and CLI new 45

Adds sandbox time-to-live (TTL) support to both the SDK and CLI, giving users control over how long a sandbox lives before it expires.

— Names SDK and CLI but no flag or parameter names given. snapshot-20260808

02 00

Python 3.10 minimum requirement breaking 40

The Python SDK now requires Python >=3.10; installations running older Python versions will break on upgrade.

— Clear version floor but no migration steps given. snapshot-20260808

└──▷ BREAKING ON UPGRADE

!
Python SDK now requires Python >=3.10; installations running older Python versions will break on upgrade.

Was this useful?

NoLabs nono

1 release · 2026-08-07 Notes Code

Sandbox any AI agent in seconds - zero setup, zero latency.

nono v0.72.0 tightens sandbox diagnostics and network enforcement on Linux, while removing two built-in profiles and changing how the external sandbox-policy mode interacts with Landlock.

└──▷ WHAT SHIPPED · 4 FEATURES 2 OF 4 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 01

External sandbox-policy mode no longer bypasses Landlock breaking 70

On Linux, the external --sandbox-policy mode no longer disables all nono sandboxing: Landlock still handles filesystem/process policy, and only TCP enforcement is delegated externally. Configurations that relied on external bypassing Landlock entirely will behave differently.

— Names the flag and describes the exact behavior change. v0.72.0

02 00

Deny-path reporting in `why` subcommand improved 65

The why subcommand now reports explicit deny paths from the sandbox policy, making it easier to diagnose why a network or filesystem access was blocked.

— Names the exact subcommand and its new diagnostic behavior. v0.72.0

03 02

Removal of openclaw and swival built-in profiles breaking 60

Removes openclaw and swival as built-in profiles; users relying on them must pull the equivalent packages from the registry, and any setup using them will break on upgrade until they do so.

— Names both removed profiles and the required workaround. v0.72.0

thinner coverage below

04 00

Static seccomp network baseline on Linux new 25

Adds a static seccomp network baseline on Linux as part of the sandbox policy enforcement.

— Named only in summary, no mechanism or usage detail given. v0.72.0

└──▷ BREAKING ON UPGRADE

!
The openclaw and swival built-in profiles have been removed; any setup using them will break on upgrade until the equivalent profiles are pulled from the registry.
!
The external --sandbox-policy mode on Linux no longer disables all nono sandboxing: Landlock still handles filesystem/process policy, and only TCP enforcement is delegated externally. Configurations that relied on external bypassing Landlock will behave differently.

Was this useful?

SecureAgentics Adrian

changes since 2026-07-09 Code

Open-source runtime AI agent security tool - monitors and controls AI agents, catching malicious tool use, prompt injection, and policy drift in real time, before the agent acts.

Adrian gains a Claude Code plugin with HITL verdict forwarding, three execution modes, and PII filtering for agent conversations.

New Claude Code plugin (integrations/claude-code/adrian_cc/agent.py) connects Claude Code hooks directly to the Adrian backend via WebSocket, sending SessionLogin + PairedEvent protobuf messages and returning allow/deny/advisory verdicts.
Supports three server-driven execution modes read from PolicySnapshot on LoginAck: MODE_ALERT (fire-and-forget, never blocks), MODE_BLOCK (waits for verdict, blocks per policy), and MODE_HITL (waits for human approval on the dashboard).
Adds HITL verdict forwarding: human approval decisions made on the Adrian dashboard are now forwarded inline to Claude Code integrations tagged with the CC source.
Adds PII filtering for the Claude Code plugin (integrations/claude-code/adrian_cc/pii/_patterns.py), detecting and redacting EMAIL, PHONE, SSN, CREDIT_CARD, IP_ADDRESS, DATE_OF_BIRTH, IBAN, PASSPORT, STREET_ADDRESS, POSTAL_CODE, DRIVER_LICENSE, and AWS_KEY categories from agent conversation content.
Adds a transcript parser (integrations/claude-code/adrian_cc/transcript.py) that reads Claude Code's JSONL transcript to extract thinking blocks, user prompts, system prompt context, and invocation metadata, with mtime-based caching to avoid re-parsing on every hook call.

5 details in Everything

Was this useful?

Cloudflare

1 release · 2026-08-08 Changelog

Cloudflare is a global network platform providing CDN, DDoS protection, DNS, security, and developer services for websites and applications.

Cloudflare's biggest change this window unifies Workers AI and AI Gateway billing and raises frontier-model rate limits, alongside a new low-overhead Kitesurf browser for agentic workloads, expanded Radar BGP APIs and AI-driven Radar tooling, and new AI Search crawling and public-endpoint options.

└──▷ WHAT SHIPPED · 8 FEATURES 0 OF 8 FEATURES most completely described first

what's the number?

Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.

depth0–40 what it does and how it works or what changed
specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point

Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.

01 00

Kitesurf browser option for Browser Run new 85

Adds a browser=kitesurf parameter to Browser Run CDP and Quick Action endpoints to opt into Kitesurf, a new stateless browser using 3–7x less CPU and memory than Chromium, aimed at agentic workloads.

Take a screenshot via Browser Run using Kitesurf to reduce CPU/memory overhead in an AI agent pipeline.

$ curl -X GET 'https://api.cloudflare.com/client/v4/accounts/<ACCOUNT_ID>/browser-run/screenshot?browser=kitesurf' \
  -H 'Authorization: Bearer <API_TOKEN>' \
  -H 'Content-Type: application/json' \
  -d '{"url": "https://example.com"}' \
  --output screenshot.png

— Exact param, endpoints and a runnable curl example changelog-20260808-37b7380e

02 00

Discover crawl mode for AI Search web sources new 85

Adds parse_type: discover for AI Search web-crawler data sources, with discover_options fields (source, limit, depth) to crawl pages beyond a sitemap by following links.

Index a site without a complete sitemap by using the discover parse type to follow links up to a set depth and page limit.

$ curl -X POST 'https://api.cloudflare.com/client/v4/accounts/<ACCOUNT_ID>/ai-search/instances' \
  -H 'Authorization: Bearer <API_TOKEN>' \
  -H 'Content-Type: application/json' \
  -d '{
    "id": "my-ai-search",
    "type": "web-crawler",
    "source": "example.com",
    "source_params": {
      "web_crawler": {
        "parse_type": "discover",
        "discover_options": { "source": "links", "limit": 5000, "depth": 3 }
      }
    }
  }'

— Named fields plus a full runnable API example changelog-20260808-37b7380e

03 00

Unified Workers AI and AI Gateway billing and routing new 80

Introduces unified billing between Workers AI and AI Gateway: setting the gateway's Workers AI billing setting to Unified Billing lets prepaid AI Gateway credits cover Workers AI inference in real time. Frontier model rate limits rise to 50 requests per minute per account per model when billed through AI Gateway credits, up from 20 via standard Workers AI billing. env.AI.run() requests can be routed through an AI Gateway by specifying gateway: { id: "<gateway-id>" }, with "default" auto-creating a gateway on first authenticated request.

— Names config keys and exact limits but no runnable example changelog-20260808-37b7380e

04 00

Public endpoints for AI Search on custom domains and namespaces new 75

Enables AI Search public endpoints on custom domains (e.g. https://search.example.com/search) with optional Cloudflare Access authentication in front of them, and adds namespace-level public endpoints (https://ns-<NAMESPACE_ENDPOINT_ID>.search.ai.cloudflare.com/search and /chat/completions) that fan out across multiple AI Search instances via ai_search_options.instance_ids.

— Exact endpoint URLs and field named, no example call changelog-20260808-37b7380e

05 00

Radar Researcher AI assistant new 60

Introduces Radar Researcher, a beta AI assistant on Cloudflare Radar with voice/text queries, interactive charts, searchable conversation history, shareable links, and an Explain with AI per-chart option.

— Describes capabilities but only points to the Radar UI changelog-20260808-37b7380e

thinner coverage below

06 00

AS-level connectivity graph and upstream widget on Radar new 55

Adds an AS-level connectivity graph and Upstream providers widget to Cloudflare Radar AS pages (e.g. AS13335), with a Show full paths toggle and an IP version selector for IPv4/IPv6.

— UI feature with a starting page but no exact steps changelog-20260808-37b7380e

07 00

cloudflare/mesh Docker image new 55

Publishes the cloudflare/mesh Docker image on Docker Hub supporting amd64 and arm64, enabling Cloudflare Mesh nodes to run in Docker Compose, Kubernetes StatefulSet, Kubernetes sidecar, and CI/CD pipelines.

— Names image and platforms but no command shown changelog-20260808-37b7380e

08 00

WebMCP support in Cloudflare Radar new 45

Adds WebMCP support to Cloudflare Radar, enabling browser-based AI agents to navigate Radar, search data, and use tools such as URL scanning and domain lookup.

— Prose description without concrete steps or interface names changelog-20260808-37b7380e

Was this useful?

The Cyber Toolchain — issue 039, August 8, 2026

On your toolchain